📄 medicine

Masseter Muscle Reorientation Patterns Following Orthognathic Surgery in Skeletal Class III Patients: A Clustering Analysis Based on Automatic Segmentation

This study utilized deep learning-based automatic segmentation and clustering analysis of masseter muscle angles in 118 skeletal Class III patients to identify two distinct postoperative reorientation patterns, revealing that specific patterns characterized by less severe deformities, smaller mandibular setbacks, and angleplasty are associated with superior long-term skeletal stability and reduced relapse.

Original authors: Longhao Fan, Yiran Jiang, Siting Chen, Yuxi Zhao, Ruoping Jiang

Published 2026-06-28

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Longhao Fan, Yiran Jiang, Siting Chen, Yuxi Zhao, Ruoping Jiang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Realigning the Jaw and the "Muscle Spring"

Imagine your jaw and the muscles that chew (specifically the masseter muscle, which is the big muscle on the side of your face) are like a complex suspension system on a car. Over years, this system has adjusted to fit your specific bone structure. If you have a "Class III" skeletal problem (often called an underbite), your jaw is positioned differently than normal, and your chewing muscles have stretched or tightened to accommodate that position.

Orthognathic surgery is like taking the car apart and moving the frame (the jawbone) to a new, straighter position. The big question the researchers asked was: What happens to the "suspension springs" (the muscles) after we move the frame? Do they snap back? Do they stretch out slowly? Or do they behave differently depending on how big a move we made?

The Experiment: Tracking 118 Patients

The researchers looked at 118 patients who had surgery to fix their underbites. They took 3D scans (like high-tech X-rays) at three specific times:

Before surgery (T0): The "before" picture.
Immediately after surgery (T1): The "freshly fixed" picture.
6 to 12 months later (T2): The "settled in" picture.

Instead of manually tracing the muscles (which is slow and prone to human error), they used a smart computer program (AI) to automatically find and measure the muscles in the scans. They then used a statistical method called clustering to see if the patients naturally fell into different groups based on how their muscles moved.

The Discovery: Two Different "Muscle Stories"

The study found that the muscles didn't all react the same way. They split the patients into two distinct groups, or "clusters," based on how the angle of their chewing muscle changed over time.

Group 1: The "Smooth Adjusters"

What happened: During the surgery, the muscle angle changed in all sorts of ways (some went one way, some the other, some stayed put). But once the surgery was done, the muscle slowly rotated in a clockwise direction as it healed.
The Analogy: Think of this like a rubber band that was stretched a little bit, then gently released. It slowly settles into a new, comfortable position without fighting back too hard.
Who they were: These patients generally had a less severe underbite to begin with. The surgeons moved their jaw back a smaller distance, and they were more likely to have a specific extra procedure called Mandibular Angleplasty (MAP) (which involves trimming the jawbone angle to reduce muscle tension).

Group 2: The "Bouncy Backers"

What happened: During the surgery, the muscle was forced to rotate clockwise (tightening up). But after the surgery, it didn't stay there. It rebounded and rotated counter-clockwise as it healed.
The Analogy: Imagine you stretch a rubber band very far and hold it tight. When you let go, it snaps back in the opposite direction. This group's muscles seemed to be under more tension during the surgery, and they "fought back" against the new position during healing.
Who they were: These patients had a more severe underbite. The surgeons had to move their jaw back much further and rotate the bone segments more aggressively. They were less likely to have the extra "tension-releasing" jaw trimming (MAP).

The Connection to Stability

The most important finding was about stability.

Group 1 (Smooth Adjusters) ended up with a more stable bite. Their jaw stayed in the new, corrected position better.
Group 2 (Bouncy Backers) experienced more relapse. This means their jaw started to drift back toward its original, pre-surgery position.

The researchers suggest that when the surgery involves moving the jaw too far or rotating it too much without releasing the muscle tension (via the MAP procedure), the muscle acts like a tight rubber band, pulling the jaw back.

Why This Matters (According to the Paper)

The paper concludes that not all patients are the same. The way the muscle reacts depends on:

How bad the underbite was to start with.
How far the jaw had to be moved.
Whether the surgeon trimmed the jaw angle to loosen the muscle.

The researchers propose that measuring the angle of the muscle (MMFHS) on a 3D scan could act like a dashboard warning light. If the muscle angle changes in a specific way (like the "Bouncy Backer" pattern), it might predict that the jaw is at higher risk of drifting back.

Summary in One Sentence

This study used AI to watch how chewing muscles react to jaw surgery and found that patients with severe underbites who had big jaw movements without muscle-relieving procedures often saw their muscles "snap back," causing their jaw to drift, while those with milder cases or extra muscle-relieving steps had smoother, more stable results.

Technical Summary: Masseter Muscle Reorientation Patterns Following Orthognathic Surgery in Skeletal Class III Patients

Problem Statement
While it is established that the masseter muscle (MM) adapts following orthognathic surgery, the specific temporal trajectories of its reorientation remain unclear. Previous studies have largely relied on small sample sizes and group-averaged comparisons between preoperative and long-term postoperative states, potentially masking divergent individual dynamic changes. Furthermore, the relationship between specific MM reorientation patterns, preoperative deformity severity, surgical variables, and postoperative skeletal stability has not been systematically explored. This study addresses the need to identify distinct patterns of MM reorientation in skeletal Class III patients and evaluate their clinical correlates and impact on stability.

Methodology
The study employed a retrospective design involving 118 adult patients with essentially symmetric skeletal Class III deformities who underwent orthognathic surgery (LeFort I osteotomy and bilateral sagittal split osteotomy [BSSO], with or without mandibular angleplasty [MAP] and genioplasty).

Data Acquisition: Cone-beam computed tomography (CBCT) or CT scans were collected at three time points: preoperative (T0), immediately postoperative (T1), and 6–12 months postoperative (T2).
Automated Segmentation: A deep learning model, previously validated for high accuracy in segmenting the MM from CBCT/CT images, was used to automatically segment the masseter muscle in all time points.
Feature Extraction: The long axis of the MM was fitted using Principal Component Analysis (PCA) on the geometric centroids of serial cross-sections. The Masseter Muscle Angle in the sagittal plane relative to the Frankfort horizontal line (MMFHS) was calculated.
Clustering Analysis: Changes in MMFHS during the intraoperative interval (T1–T0) and the postoperative interval (T2–T1) were used as clustering features. K-means clustering analysis was performed, with the optimal number of clusters ( $k=2$ ) determined via the Elbow method. Stability was assessed using bootstrap resampling (2,000 iterations).
Statistical Analysis: Between-cluster comparisons utilized t-tests, Mann-Whitney U tests, Chi-square tests, or Fisher's exact tests. Analysis of covariance (ANCOVA) was used to compare postoperative skeletal stability outcomes (T2) and relapse (T2–T1), adjusting for baseline ANB, mandibular setback, and specific rotational/vertical displacements of the proximal (PS) and distal segments (DS).

Key Contributions

Novel Methodological Approach: This is the first study to utilize an automated deep learning segmentation model for large-scale, three-dimensional quantitative analysis of MM reorientation across preoperative, immediate postoperative, and follow-up stages in a cohort of 118 patients.
Identification of Heterogeneous Patterns: The study moves beyond group averages to identify two distinct, stable reorientation trajectories of the MMFHS using unsupervised clustering.
Clinical Correlation: The study links specific reorientation patterns to preoperative deformity severity, surgical movement magnitudes (setback and rotation), and the performance of mandibular angleplasty (MAP).
Stability Association: It establishes an association between MMFHS reorientation patterns and postoperative skeletal stability, specifically anterior mandibular relapse, independent of standard surgical variables.

Results
Two distinct reorientation patterns were identified:

Cluster 1 (n=56): Characterized by variable intraoperative changes (clockwise, counterclockwise, or stable) followed by a consistent clockwise rotation postoperatively.
Cluster 2 (n=62): Characterized by clockwise rotation intraoperatively followed by a counterclockwise rebound postoperatively.

Comparative Findings:

Preoperative Characteristics: Cluster 1 patients had less severe skeletal Class III deformities (larger ANB, smaller SNB) compared to Cluster 2.
Surgical Variables: Cluster 2 underwent significantly larger mandibular setbacks (8.91 mm vs. 6.33 mm) and greater clockwise rotation of the proximal segment (4.00° vs. 2.68°) and distal segment compared to Cluster 1.
MAP Utilization: The proportion of patients undergoing bilateral MAP was significantly higher in Cluster 1 (30.4%) than in Cluster 2 (8.1%).
Postoperative Stability: After adjusting for baseline ANB, mandibular setback, and segmental rotations, Cluster 1 demonstrated significantly larger postoperative ANB angles and significantly less anterior mandibular relapse (-1.43 mm) compared to Cluster 2 (-2.55 mm).

Significance and Claims
The authors claim that the reorientation of the MMFHS following orthognathic surgery is heterogeneous rather than uniform. The study suggests that these distinct patterns are associated with the severity of the preoperative deformity, the magnitude of surgical skeletal movements, and the performance of MAP.

The paper posits that the counterclockwise rebound observed in Cluster 2 may be a response to higher postoperative soft-tissue tension resulting from larger skeletal displacements and PS rotations. Conversely, the clockwise rotation in Cluster 1 may represent a favorable adaptive realignment toward a functional phenotype similar to skeletal Class I.

The authors conclude that managing masticatory soft-tissue tension and proximal segment positioning is critical during surgical planning. They propose that MMFHS may serve as a potential imaging marker for postoperative muscle adaptation and skeletal stability. However, the authors explicitly state that these findings are exploratory and that the observed relationships are correlations rather than causal, necessitating further validation in larger, multicenter cohorts with longer follow-up periods and functional data (e.g., electromyography, bite force).